Image Similarity / Dedup
Compute pHash and find duplicates / near-duplicates in a folder.
- Image
- Account required
- Server processed
Add your files
Run Image Similarity / Dedup
Processed on our servers. Uploaded over TLS and deleted an hour later — 24 hours if you are signed in.
Three steps
How to use Image Similarity / Dedup
- Step 01
Pick your input
Drag in one or more image/*, or drop a single ZIP — we'll unpack it for you.
- Step 02
Configure & run
Click Run — there is nothing to configure. We process the file on our own infrastructure — no third-party APIs in the path.
- Step 03
Download the result
A JSON document is downloadable below.
Server-processed files are encrypted at rest with a per-tenant key and deleted an hour after upload — 24 hours if you are signed in. We do not use third-party processing APIs.
About image similarity / dedup
Compute pHash and find duplicates / near-duplicates in a folder. It sits in the image shelf alongside 168 other image tools, needs no account, and keeps the same behaviour whether you open it here or call it through the REST API.
It takes any image you give it and gives back JSON. It uses a model we host ourselves, not a third-party API. The work happens on our servers, and the upload is encrypted per file and deleted an hour later — 24 hours if you are signed in. The same operation is available programmatically at POST /v1/image/image-similarity.
Image Similarity / Dedup questions
Where does Image Similarity / Dedup process my file?
On our own servers. This tool needs more than a browser tab can provide, so the file is uploaded over TLS, processed on our infrastructure — no third-party processing APIs are in the path — and both the input and the result are deleted within an hour.
What files can I use with Image Similarity / Dedup?
Any image file your browser can read. The limit is 100 MB per file.
Can I run Image Similarity / Dedup on several files at once?
Yes — select or drop as many as you need and they are handled together.
Do I need an account to use Image Similarity / Dedup?
No. Image Similarity / Dedup works without signing in, and that is not a trial — it is how the free tier works. An account only matters for things that need one to exist at all: larger server jobs against your quota, API keys, saved workflows and billing.
Can I run Image Similarity / Dedup from my own code?
Yes. Every tool on the site is reachable from the REST API with the same slug — `image-similarity` — and the same options as this page, so anything you can do here you can automate against your own files.